NeST: nested hierarchical structure identification in spatial transcriptomic data.

Area of Science:

• Genomics
• Bioinformatics
• Systems Biology

Background:

• Spatial gene expression data reveals regional gene enrichment/depletion.
• Existing segmentation methods struggle with nested structures and scale.
• Discovering hierarchical patterns in gene expression is crucial for understanding tissue organization.

Purpose of the Study:

• Introduce NeST for extracting hierarchical spatial structures in transcriptomic data.
• Enable the discovery of nested subregions with distinct expression patterns.
• Facilitate de novo analysis of cell-cell interactions using ligand-receptor data.

Main Methods:

• NeST identifies "coexpression hotspots" representing localized spatial coexpression of gene sets.
• The method analyzes structure across any spatial scale and gene subset.
• NeST performs 3D spatial analysis of ligand-receptor interactions.

Main Results:

• NeST successfully reveals nested biological structures in various spatial transcriptomic datasets.
• Coexpression hotspots provide explainable insights into spatial gene organization.
• The tool identifies active areas of cell-cell interaction without prior grouping.

Conclusions:

• NeST uncovers a novel level of biological organization in spatial transcriptomic data.
• The method offers a scalable and explainable approach to spatial analysis.
• NeST enhances the understanding of tissue architecture and cell communication.